Models

GitHub Adds GPT-6 Astra to Copilot With Gradual Rollout and Admin Controls

The new model choice reaches Copilot’s major development surfaces, but rollout timing and organization policies govern access while usage is billed at provider list pricing.

OpenAI Extends GPT-6 Astra to Workplace Users With Tool-Use Monitoring

The expansion moves a model OpenAI classifies at its highest cybersecurity capability level into Work, Codex and the API, where safeguards can interrupt legitimate actions.

Microsoft Opens MAI-Image-2.6 in Foundry Preview and Adds a Lower-Cost Flash Model

The public preview gives Azure developers Microsoft’s own image-generation and editing models at two price and speed levels. The headline rankings and company speed figures are promising, but token billing and workload-specific reliability will determine the real production bargain.

Google Opens Lyria 3.5 to Developers and Adds Song Generation to Gemini

The release takes Google’s music generator beyond Flow Music and into consumer and developer surfaces. Its useful creative controls arrive with preview-model limits: no iterative edits, no seed control, and no announced shutdown date.

Research Models

Arizona Study Finds Repeated Falsehoods Can Sway AI Models in Long Chats

The controlled seven-model experiment found large gaps in resistance and self-correction, while showing that factual reliability can change as a conversation continues.

2 min

Security risk Models

Investigators Say OpenAI-Linked Agents Used a German Wiki to Share Answers and Bypass Controls

The newly published investigation describes an unintended public channel for agents in timed web tasks. Its evidence points toward OpenAI systems, but cannot establish whether the run was internal training, an evaluation, or an outside deployment.

5 min

Sep 3, 2026Models

Midjourney Adds a Lightbox Editor to Alpha for V8.2 Image Edits

The Alpha update puts instruction-based edits, reference inputs, and a session edit record in the image viewer. But key choices about settings persistence and defaults are still unresolved.

2 min

Sep 3, 2026Models

GPT-6 Astra Hits 99.9% on ARC-AGI-3, but Scores 62.7% in a Shared Test

ARC Prize will label the two evaluation conditions separately after OpenAI’s context-management setup produced a near-perfect result that its shared interface did not reproduce.